ESRI generally doesn't post the equations that they use, but may give reference to cited papers and/or books. You can be assured that the IDW equation implemented is not of their own making but has a long history of implementation.
ESRI generally doesn't post the equations that they use, but may give reference to cited papers and/or books. You can be assured that the IDW equation implemented is not of their own making but has a long history of implementation.
thanks for the contribution.
I tested the result of the IDW and got confused about some resulted values. just i wanted to know if the equation used the same as the one i posted.
best
Jamal
---------------------------------------- Jamal Numan Geomolg Geoportal for Spatial Information Ramallah, West Bank, Palestine
This is quite helpful. Just I was trying to calculate some values manually applying the equation I�??ve provided in my previous post and got confused about some results. I wanted to make sure about the math used in the ArcGIS to calculate the IDW.
I�??m wondering why ESRI doesn�??t include on the help the math they use in such calculations. Another issue that I�??m confronting for the time being is the equation used to calculate the kernel density. How could I obtain the formula used in the such computations?
Best
Jamal
---------------------------------------- Jamal Numan Geomolg Geoportal for Spatial Information Ramallah, West Bank, Palestine
It is often better to cite a reference rather than copy from it. This will give you more insight into the inner workings and assumptions around these equations.
Which is what we have done for IDW and Kernel Density.
It is often better to cite a reference rather than copy from it. This will give you more insight into the inner workings and assumptions around these equations.
Which is what we have done for IDW and Kernel Density.
Steve
Many thanks Steve.
best
Jamal
---------------------------------------- Jamal Numan Geomolg Geoportal for Spatial Information Ramallah, West Bank, Palestine
No argument there. But sometimes the coder has a different understanding of the references they are using than do other readers. Thus, as documentation of the software, the developer can do very well by describing how they interpreted the equations and implemented in them in the software. This closes the loop from equation to software back to equation. Furthermore, by not forcing software users to reference the literature (which often is academic papers not freely available), this makes explicit documentation available to all users, not just the privileged ones with good library access.
A recent (nice) example of the value of explicit documentation compared to a mere reference appears at on the StackExchange GIS Q&A site, where a user of open source GIS software found a problem not by testing the software, but by seeing something strange in the documentation (which, because this is OSS, includes the code itself, which is not an option with ArcGIS). The code was based on good references but mis-interpreted them (and made a few additional mistakes along the way...).
I, for one, have appreciated the accessibility of the Python source code in some of the ArcToolboxes, because it goes a long way to showing that these tools are doing what I expect and are doing it correctly. Otherwise, when working on important projects (and all projects for clients are important!), I have to test the software extensively, because the documentation tends to be non-technical and vague.
This is exactly what I needed to know. Including the equations used in the computation is core issue in my opinion. This lets know where we are exactly and based on what the results were computed. Indicating the reference without full details on what exactly have been implemented (particularly where there are definite equations) doesn�??t help the user to be confident about the entire process.
Best
Jamal
---------------------------------------- Jamal Numan Geomolg Geoportal for Spatial Information Ramallah, West Bank, Palestine
the old post about this mentiones the following from Bill Huber:
start quote================= The very name, "inverse distance weighted," practically describes the formula.
The objective is to interpolate a value Z at a location P, given values Z(1), ..., Z(n) at distinct other locations P(1), ..., P(n), respectively. The interpolated value is a weighted average of the Z's. That is, it develops a sequence of numbers (the "weights") w(1), ..., w(n) and forms
Z = [w(1)*Z(1) + ... + w(n)*Z(n)]/[w(1)+...+w(n)]
The division by the sum of the weights is what makes this a true average.
A formula like this applies to many interpolators, such as Kriging; the issue is how the weights are computed. With IDW it's simple and straightforward: you select a power 'p', usually around 2, and compute
w(i) = 1/Distance(P, P(i))^p.
That is, the weights are inversely proportional to the power 'p' of the distances between the point of interpolation and the data locations. That's all there is to it.
end quote ===============
however: what distance? if it is in kilometers the surface is smoother than when it is in meters.
